1598
Catalytic Chan–Lam coupling using a ‘tube-in-tube’ reactor todeliver molecular oxygen as an oxidantCarl J. Mallia1, Paul M. Burton2, Alexander M. R. Smith2, Gary C. Walter2
and Ian R. Baxendale*1
Full Research Paper Open Access
Address:1Department of Chemistry, Durham University, South Road, Durham,DH1 3LE, United Kingdom and 2Syngenta, Jealott's Hill InternationalResearch Centre, Bracknell, Berkshire, RG42 6EY, United Kingdom
Email:Ian R. Baxendale* - [email protected]
* Corresponding author
Keywords:Chan–Lam coupling; flow chemistry; gases in flow; oxygen;“tube-in-tube”
Beilstein J. Org. Chem. 2016, 12, 1598–1607.doi:10.3762/bjoc.12.156
Received: 14 May 2016Accepted: 13 July 2016Published: 26 July 2016
This article is part of the Thematic Series "Automated chemicalsynthesis".
Associate Editor: A. Kirschning
© 2016 Mallia et al.; licensee Beilstein-Institut.License and terms: see end of document.
AbstractA flow system to perform Chan–Lam coupling reactions of various amines and arylboronic acids has been realised employing mo-
lecular oxygen as an oxidant for the re-oxidation of the copper catalyst enabling a catalytic process. A tube-in-tube gas reactor has
been used to simplify the delivery of the oxygen accelerating the optimisation phase and allowing easy access to elevated pressures.
A small exemplification library of heteroaromatic products has been prepared and the process has been shown to be robust over ex-
tended reaction times.
1598
IntroductionThe functionalisation of aromatic and aliphatic amines has
received considerable attention due to the number of biological-
ly active compounds represented by these classes. For this
reason different synthetic methods for C–N bond formation
have been developed (Scheme 1) over the years with the general
goal to overcome the shortcomings of the original Ullman [1]
and Goldberg [2] methods relating to the harsh reaction condi-
tions they employ. After a closer look at the work of Mitiga [3]
on the Stille coupling reactions, Hartwig [4] and Buchwald [5]
independently proposed a catalytic mechanism and later re-
ported a tin free aryl–amine coupling reaction [6,7]. This major
breakthrough made the C–N coupling reaction accessible to a
wide range of substrates, including anilines, which did not react
very well with the previous conditions. However, despite the
improvements achieved with the Buchwald–Hartwig coupling,
limitations such as sensitivity to air and moisture, functional
group tolerance and the high cost of palladium, reignited the
search for an improved method.
In 1998, the groups of Chan [8], Evans [9] and Lam [10] inde-
pendently reported upon mild methods for C(aryl)–N and
C(aryl)–O coupling reactions. Their methods made use of stoi-
chiometric amounts of copper(II) acetate as the catalyst and
boronic acids as the aryl donors. In the presence of a base, the
Beilstein J. Org. Chem. 2016, 12, 1598–1607.
1599
Scheme 1: Comparison of early C–N and C–O coupling reactions.
coupling could be performed at room temperature. These reac-
tions were subsequently shown to work with a large number of
nucleophiles and tolerated a variety of substrates, making the
process one of the most efficient ways for C–N/O coupling
[11]. Several modifications of the Chan–Lam reaction have
been reported, expanding its scope and it has since been used to
synthesise several biologically active compounds [11,12].
In 2009 the groups of Stevens and van der Eycken reported on
the Chan–Lam reaction as a continuous flow protocol using
copper(II) acetate (1.0 equiv), pyridine (2.0 equiv) and triethyl-
amine (1.0 equiv) in dichloromethane [13]. Generally, when
using anilines or phenols as the nucleophilic partner, moderate
to good yields were obtained (56–71% yields, 9 examples).
More recently the Tranmer group reported the use of a copper-
filled column as a catalyst with TEMPO as the co-oxidant in
acetonitrile (acetic acid additive) with moderate to good yields
of the coupled products being obtained (25–79% yields, 16 ex-
amples) [14]. The use of a copper tubing which serves as both
the reactor and the catalyst with tert-butyl peroxybenzoate as
the oxidant in acetonitrile was also described but was outper-
formed by the copper filled column system. Although the use of
elemental copper is potentially an improvement on the use of
stiochiometric copper(II) acetate in continuous flow, the use of
TEMPO or tert-butyl peroxybenzoate as a co-oxidant intro-
duces waste. Employing oxygen gas as an oxidant is preferred
as it is cheap, renewable and environmentally benign. We there-
fore set out to develop a more atom economical way of
catalysing the Chan–Lam reaction using a sub-stoichiometric
amount of copper and oxygen gas as the oxidant.
The use of oxygen provides the necessary oxidant to reoxidise
the Cu(I) that forms after the C–N reductive elimination back to
Cu(II), allowing for sub-stoichiometric amounts of copper cata-
lyst to be used [15,16]. Based upon our previous experience of
using the reverse “tube-in-tube” reactor with other gases, it was
decided that oxygen would be delivered via this reactor set-up
(Figure 1).
Results and DiscussionIn our initial screening, four different organic solvents with
good oxygen absorption were investigated (toluene, dichloro-
methane, acetonitrile and ethyl acetate), however, Cu(OAc)2
was only completely soluble in dichloromethane. Consequently
dichloromethane was used as the reaction solvent. Unfortu-
nately pumping dichloromethane through the HPLC pumps,
used as part of the flow system, initially presented some issues.
This was mainly due to cavitation which occurred just before
the pump inlet, attributed to the shear forces present, causing
outgassing (air). These bubbles, if allowed to enter the system
Beilstein J. Org. Chem. 2016, 12, 1598–1607.
1600
Figure 1: General flow scheme for catalytic Chan–Lam reaction.
disturbed the flow (or impaired the pump), resulting in unstable
flow. The problem was solved when the dichloromoethane used
was sonicated (30 min of sonication per 500 mL of solvent)
prior to use, it was then maintained under positive pressure at
the inlet throughout the experiment (N2 balloon was used for
the positive pressure).
In an effort to identify the optimum conditions for the reaction
process, the amount of copper catalyst and the oxygen pressure
were studied (Table 1).
A set of control experiments with no oxygen was run and the
amount of copper acetate catalyst was lowered from 1 to
0.25 equiv (entries 1–3, Table 1). As anticipated, with no
oxidant to reoxidize the catalyst, the yield of 19 dropped in
proportion to the amount of catalyst used. Next, whilst main-
taining the amount of copper acetate (0.5 equiv), the effect of
the oxygen pressure on conversion was investigated (entries
4–8, Table 1). A general increase in the yield of 19 was ob-
tained on going from atmospheric to 10 bar after which a slight
decrease in yield was encountered at higher pressures
(Figure 2). This same decrease in yield was also observed when
going from 10 bar to 12 bar of oxygen using 0.25 equiv of
copper acetate (entries 9 and 10, Table 1).
When the amount of copper acetate was reduced to 0.1 equiv a
drastic decrease in yield was observed indicating that the TOF
of the catalyst prevented achievement of good yields within the
time limits (residence time) of the flow reactor (entry 11,
Table 1). A decrease in yield was also observed when the
amount of boronic acid used was decreased to 1.4 equiv and
1.1 equiv, respectively (entries 12 and 13, Table 1). Changing
the temperature from 20 °C to 50 °C did not greatly affect the
yields obtained, with 40 °C giving the most promising result
(entries 14–16, Table 1). However, it was observed that less
particulate matter was formed in the reactor when higher tem-
Beilstein J. Org. Chem. 2016, 12, 1598–1607.
1601
Table 1: Optimisation of the Chan–Lam reaction in continuous flow.
Entry Cu(OAc)2 (equiv) Boronic acid (equiv) Temperature (°C) O2 pressure (bar) NMR conversion (%)a
1 1.00 1.6 20 0 66
2 0.50 1.6 20 0 48
3 0.25 1.6 20 0 25
4 0.50 1.6 20 4 81
5 0.50 1.6 20 8 85
6 0.50 1.6 20 10 97
7 0.50 1.6 20 12 85
8 0.50 1.6 20 14 83
9 0.25 1.6 20 10 94
10 0.25 1.6 20 12 87
11 0.10 1.6 20 10 50
12 0.25 1.4 20 10 56
13 0.25 1.1 20 10 48
14 0.25 1.6 30 10 87
15 0.25 1.6 40 10 95
16 0.25 1.6 50 10 88
17b 0.25 1.6 40 10 93
18c 0.25 1.6 40 10 76aYields calculated using 1,3,5-trimethoxybenzene as an internal NMR standard and represents the average of two runs. b1.5 equiv of pyridine,c0.5 equiv of pyridine.
peratures were used (40 and 50 °C), which helps in avoiding
possible reactor blockages. Finally, the amount of pyridine
added was also studied. Decreasing the amount of pyridine
(0.5 equiv, entry 18, Table 1) resulted in a lower yield (76%)
while increasing the amount of pyridine (1.5 equiv, entry 18,
Table 1) did not produce any noticeable change in the yield
(93%). This indicates that the pyridine plays an important role
in this coupling reaction which could be both due to its effect as
a ligand and/or its solubility enhancement of the copper acetate.
The amount of triethylamine was not varied as its quantity was
required to ensure the boronic acid remained soluble in the
dichloromethane solvent.
To determine the time needed to reach steady state in the
reactor, samples were periodically collected (every 2 min via an
autosampler) and analysed by 1H NMR spectroscopy using
1,3,5-trimethoxybenzene as an internal standard. As expected,
the product started eluting after 120 min which corresponds
Beilstein J. Org. Chem. 2016, 12, 1598–1607.
1602
Figure 3: Comparison of 1H NMR spectra of non-purified (top) and QP-DMA purified (bottom) continuous flow synthesis of compound 20.
Figure 2: Observed trend for the effect of changing oxygen pressureon the NMR yield of 19.
with the theoretical residence time. A lower yield was initially
obtained for 120 min (85% yield) which then rapidly increased
to 98% yield at 125 min. The yield then stabilised from 135 min
at 96% indicating steady state was achieved.
As it had been determined that the amount of arylboronic acid
excess could not be lowered (entries 12 and 13, Table 1), the
use of a polymer supported scavenger was tested in an effort to
sequester the excess boronic acid. A column of QP-DMA, a
polymer-supported tertiary amine base, was placed in-line after
the “tube-in-tube” reactor (Figure 1). It was found that this was
sufficient to remove the majority of boronic acid without
affecting the yield of the product (Figure 3). Ultimately as the
Beilstein J. Org. Chem. 2016, 12, 1598–1607.
1603
Scheme 2: Scope of the catalytic Chan–Lam reaction in continuous flow.
products were required for biological screening they were still
purified by column chromatography, however, the reduction of
the boronic acid excess made the chromatography far easier.
Reaction scoping and library preparationUsing the optimised conditions determined for the synthesis of
compound 19, a small library was prepared to demonstrate the
scope of the reaction conditions. Excellent isolated yields were
obtained when anilines were used as the nucleophilic partner
with both 4-methoxyphenylboronic acid (90% yield of 21) and
phenylboronic acid (92% yield of 22) as the aryl donors
(Scheme 2). Phenylboronic acid also gave a moderate isolated
yield when coupled with 3-amino-5-bromopyridine as the
nucleophile (50% yield of 23, Scheme 2) and a good isolated
yield with the electron withdrawing 4-chloroaniline (71% yield
of 24, Scheme 2). Using L-tyrosine methyl ester as the nucleo-
phile with phenylboronic acid, unfortunately, gave a poor isolat-
ed yield of 26% and also underwent some epimerisation (25,
53% ee determined by chiral HPLC, Scheme 2). Additionally, a
small amount of the product (25) reacted further with phenyl-
boronic acid through the phenol to give 26 in 3% isolated yield.
In the case of L-leucine methyl ester an isolated yield of 60%
was realised, but this substrate also underwent partial epimeri-
sation (27, 71% ee determined by chiral HPLC, Scheme 2).
Using N-heterocyclic substrates as the nucleophilic partner with
a range of different phenylboronic acids generally gave good
isolated yields (19, 20, 28–35, Scheme 2). Using a pyradizine as
a nucleophilic partner an 81% yield was obtained for the forma-
tion of 20. However, using 3,4-dimethyl-1H-1,2,4-triazol-
5(4H)-one (39), which was synthesised using a literature proce-
dure [17,18] (Scheme 3), with 3,4-dimethoxyphenylboronic
acid gave a lower yield of 26% (28, Scheme 2). It is not yet
clear as to why such a low conversion and isolated yield was
Beilstein J. Org. Chem. 2016, 12, 1598–1607.
1604
Scheme 3: Syntheses of substrate 39.
Figure 4: NOESY NMR spectrum for 30 with the characteristic NOESY signal encircled.
obtained although the reduced nucleophilicity and higher poten-
tial for coordination of the triazole to the copper catalyst might
inhibit catalyst turnover and account for this.
Alternatively, using 3-phenyl-1H-pyrazole (18) as the nucleo-
phile with a number of different phenylboronic acids gave mod-
erate to good yields (38–82% yields). In general electron-rich
phenylboronic acids (19, 29–32, Scheme 2) gave better yields
than electron poor ones (33–35, Scheme 2). This is probably
due to the more favourable thermodynamics with an increase in
the electropositive nature of boron, which in turn increases the
rate of the transmetallation step. Changing the group at the
4-position of the phenylboronic acid gave good yields for both
electron-rich (19, 79% yield) and electron-poor (33, 76% yield)
phenylboronic acids. On the other hand changing the group at
the 3-position of the phenylboronic acid gave good yields for
electron-rich (30 and 32, 77% and 82% yields, respectively) but
only a moderate yield of 40% for electron-poor (34) phenyl-
boronic acids. Lower yields were also encountered for both
electron-rich (65% yield) and electron-poor (38% yield)
2-substituted phenylboronic acids, most likely due to steric
factors (31 and 35, Scheme 2).
It is noteworthy that for all of the 3-phenyl-1H-pyrazole
couplings, only the 1,3-disubsituted pyrazole products were ob-
tained with no 1,5-disubsituted isomers being detected. The
regioselectivity of the 1,3-disubsituted pyrazoles was con-
firmed by NOESY NMR experiments (30, 33 and 35, Figures
4–6) as well as comparison to known published data. In addi-
tion, an X-ray crystal structure for compound 33 was obtained
and the connectivity confirmed. It was noted that several exam-
ples of literature reported cases where mixtures of regioisomers
had been obtained were wrongly assigned.
The process described does have certain limitations. For certain
nucleophilic substrates no products were obtained when C–N
Beilstein J. Org. Chem. 2016, 12, 1598–1607.
1605
Figure 5: NOESY NMR spectrum for 33 with the characteristic NOESY signal encircled.
Figure 6: NOESY NMR spectrum for 35 with the characteristic NOESY signal encircled.
Beilstein J. Org. Chem. 2016, 12, 1598–1607.
1606
Figure 7: Substrates that gave no products in flow.
Scheme 4: Scale-up procedure for 19.
coupling with 4-methoxyphenylboronic acid (17) was attempted
(Figure 7). In the case of substrate 40 precipitation occurred as
soon as the two solutions came into contact at the T-piece
mixer, which was probably due to strong coordination to the
copper acetate by the imidazole ring. This made running this
reaction problematic in flow due to the occurrence of reactor
blocking. Other substrates proved unreactive. In the case of
starting materials 41–43 the reduced nucleophilicity of these
substrates might account for the lack of conversion. By compar-
ison, all three substrates (41–43) also failed to react under batch
conditions using 2 equiv of Cu(OAc)2, 2 equiv of NEt3 and
1 equiv of pyridine at 40 °C for 48 h confirming their low reac-
tivity.
Reaction scalingFinally, the robustness of the process and potential for scala-
bility of the general reaction conditions was demonstrated by
the synthesis of 19 at a 10 mmol scale, a factor of fourteen
times the original 0.7 mmol test reaction (Scheme 4). A slightly
improved isolated yield (81%) was obtained for the larger scale
experiment when compared to the 79% isolated yield obtained
for the shorter run experiment. The consistency of the yields ob-
tained indicates that the process is robust and without modifica-
tion can reliably deliver 0.216 g h−1 of 19 at 81% isolated yield.
ConclusionThe use of flow chemistry for the C–N coupling through a cata-
lytic Chan–Lam reaction has allowed for a safe and efficient
introduction of oxygen through a reverse “tube-in-tube” reactor.
Optimisation of the reaction conditions allowed for a scalable
and efficient way for the continuous synthesis of a number of
functionalised aromatic and aliphatic amines including a num-
ber of 1,3-disubstituted pyrazoles which were selectively ob-
tained over the regioisomeric 1,5-disubstituted products. When
compared to other published protocols it is clear that the use of
sub-stoichiometric amounts of the copper catalysts presents an
advantage over the stoichiometric amount used in the original
flow studies [13]. Additionally, the use of oxygen as the oxidant
offers improved atom economy over the use of systems such as
TEMPO and tert-butyl peroxybenzoate [14]. We believe this
Beilstein J. Org. Chem. 2016, 12, 1598–1607.
1607
approach therefore present several opportunities for laboratory
chemists to utilise this valuable C/N coupling methodology.
ExperimentalWarning: Oxygen is a highly flammable gasand all reactions were carried out in wellventilated fume cupboardsFor the flow process, 0.781 mmol of the amine was dissolved in
5.5 mL of dichloromethane followed by 1.25 mmol of the
boronic acid and NEt3 (0.039 g, 54 µL, 0.391 mmol). Another
solution containing Cu(OAc)2·H2O (0.195 mmol, 0.25 equiv),
NEt3 (0.039 g, 54 µL, 0.391 mmol) and pyridine (0.062 g,
63 µL, 0.781 mmol) in 5.5 mL of dichloromethane was also
prepared. The two solutions were separately introduced in a
5 mL loop as shown in Table 1. The pumps were each set at
0.125 mL/min to achieve a residence time of 2 h. Two reverse
“tube-in-tube” reactors (supplied by Vapourtec) were used in
series to achieve a combined reactor volume of 30 mL which
were heated at 40 ºC. The reaction mixture was then passed
through an Omnifit column (r = 0.33 cm, h = 10.00 cm) filled
with QP-DMA followed by a back pressure regulator (175 psi).
The crude reaction mixture was then passed through a plug of
silica to remove most of the excess copper present and the
organic solvent from eluent evaporated under reduced pressure.
The resultant crude material was then purified using flash chro-
matography.
Supporting InformationSupporting Information File 1Experimental procedures and characterization data for all
new compounds.
[http://www.beilstein-journals.org/bjoc/content/
supplementary/1860-5397-12-156-S1.pdf]
AcknowledgementsWe would like to acknowledge the funding and support from
the Royal Society (to I.R.B.; UF130576) and EPSRC/Syngenta
(to C.J.M.; Grant No. EPSRC 000228396) that has enabled this
work to be undertaken. Furthermore, we are grateful to Dr A.
Batsanov (Durham University, Department of Chemistry) for
solving the X-ray structure.
References1. Ullmann, F.; Sponagel, P. Ber. Dtsch. Chem. Ges. 1905, 38,
2211–2212. doi:10.1002/cber.1905038021762. Goldberg, I. Ber. Dtsch. Chem. Ges. 1906, 39, 1691–1692.
doi:10.1002/cber.19060390298
3. Kosugi, M.; Kameyama, M.; Migita, T. Chem. Lett. 1983, 12, 927–928.doi:10.1246/cl.1983.927
4. Paul, F.; Patt, J.; Hartwig, J. F. J. Am. Chem. Soc. 1994, 116,5969–5970. doi:10.1021/ja00092a058
5. Guram, A. S.; Buchwald, S. L. J. Am. Chem. Soc. 1994, 116,7901–7902. doi:10.1021/ja00096a059
6. Guram, A. S.; Rennels, R. A.; Buchwald, S. L.Angew. Chem., Int. Ed. Engl. 1995, 34, 1348–1350.doi:10.1002/anie.199513481
7. Louie, J.; Hartwig, J. F. Tetrahedron Lett. 1995, 36, 3609–3612.doi:10.1016/0040-4039(95)00605-C
8. Chan, D. M. T.; Monaco, K. L.; Wang, R.-P.; Winters, M. P.Tetrahedron Lett. 1998, 39, 2933–2936.doi:10.1016/S0040-4039(98)00503-6
9. Evans, D. A.; Katz, J. L.; West, T. R. Tetrahedron Lett. 1998, 39,2937–2940. doi:10.1016/S0040-4039(98)00502-4
10. Lam, P. Y. S.; Clark, C. G.; Saubern, S.; Adams, J.; Winters, M. P.;Chan, D. M. T.; Combs, A. Tetrahedron Lett. 1998, 39, 2941–2944.doi:10.1016/S0040-4039(98)00504-8
11. Rao, K. S.; Wu, T.-S. Tetrahedron 2012, 68, 7735–7754.doi:10.1016/j.tet.2012.06.015
12. Fischer, C.; Koenig, B. Beilstein J. Org. Chem. 2011, 7, 59–74.doi:10.3762/bjoc.7.10
13. Singh, B. K.; Stevens, C. V.; Acke, D. R. J.; Parmar, V. S.;Van der Eycken, E. V. Tetrahedron Lett. 2009, 50, 15–18.doi:10.1016/j.tetlet.2008.09.159
14. Bao, J.; Tranmer, G. K. Tetrahedron Lett. 2016, 57, 654–657.doi:10.1016/j.tetlet.2015.12.107
15. King, A. E.; Brunold, T. C.; Stahl, S. S. J. Am. Chem. Soc. 2009, 131,5044–5045. doi:10.1021/ja9006657
16. King, A. E.; Huffman, L. M.; Casitas, A.; Costas, M.; Ribas, X.;Stahl, S. S. J. Am. Chem. Soc. 2010, 132, 12068–12073.doi:10.1021/ja1045378
17. Kahveci, B. Molecules 2005, 10, 376–382. doi:10.3390/1002037618. Ün, R.; Îkizler, A. Chim. Acta Turc. 1975, 3, 113–132.
License and TermsThis is an Open Access article under the terms of the
Creative Commons Attribution License
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
The definitive version of this article is the electronic one
which can be found at:
doi:10.3762/bjoc.12.156